def GetAsrResult(self):
#本地文件方式请求
try:
cred = credential.Credential(self.SecretID, self.SecretKey)
httpProfile = HttpProfile()
httpProfile.endpoint = self.Url
clientProfile = ClientProfile()
clientProfile.httpProfile = httpProfile
clientProfile.signMethod = "TC3-HMAC-SHA256"
client = asr_client.AsrClient(cred, "ap-shanghai", clientProfile)
#读取文件以及 base64
with open(self.AudioFile, "rb") as f:
audio = f.read()
if self.Format == 'pcm':
audio = utils.PcmToWav(audio)
if sys.version_info[0] == 2:
content = base64.b64encode(audio)
else:
content = base64.b64encode(audio).decode('utf-8')
#发送请求
req = models.SentenceRecognitionRequest()
params = {"ProjectId": 0, "SubServiceType": 2, "SourceType": 1, "UsrAudioKey": "session-123"}
req._deserialize(params)
req.DataLen = len(content)
req.Data = content
req.EngSerViceType = "16k_zh"
req.VoiceFormat = "wav"
req.FilterPunc = 2 #是否过滤标点符号（目前支持中文普通话引擎）。 0：不过滤，1：过滤句末标点，2：过滤所有标点。默认为0。
req.ConvertNumMode = 0 #是否进行阿拉伯数字智能转换。0：不转换，直接输出中文数字，1：根据场景智能转换为阿拉伯数字。默认值为1
resp = client.SentenceRecognition(req)
logger.info("file:%s, result:%s, sid:%s", self.AudioFile, resp.Result, resp.RequestId)
return resp.Result
except TencentCloudSDKException as err:
logger.error("code:%s, message:%s, sid:%s", err.code, err.message, err.requestId)
return ""
def stt(self, audio_file):
try:
fwave = open(audio_file, mode='rb')
data = fwave.read()
dataLen = len(data)
base64Wav = base64.b64encode(data)
fwave.close()
params = '{"ProjectId":0,"SubServiceType":2,"EngSerViceType":"16k","SourceType":1,"VoiceFormat":"mp3","UsrAudioKey":"session-123", ' + '"Data":"' + str(
base64Wav, 'utf-8') + '", "DataLen":' + str(dataLen) + '}'
cred = credential.Credential(self.app_id, self.app_key)
httpProfile = HttpProfile()
httpProfile.endpoint = self.api_url
clientProfile = ClientProfile()
clientProfile.httpProfile = httpProfile
client = asr_client.AsrClient(cred, "ap-chengdu", clientProfile)
req = models.SentenceRecognitionRequest()
req.from_json_string(params)
res = client.SentenceRecognition(req)
resp = json.loads(res.to_json_string())
return resp['Result']
except TencentCloudSDKException as err:
ASRServerError(err, "404")
def recognize_wav(wav_path, show_detail=True):
# 通过本地语音上传方式调用
try:
# 重要：<Your SecretId>、<Your SecretKey>需要替换成用户自己的账号信息
# 请参考接口说明中的使用步骤1进行获取。
cred = credential.Credential(tencent_secret_id, tencent_secret_key)
httpProfile = HttpProfile()
httpProfile.endpoint = "asr.tencentcloudapi.com"
clientProfile = ClientProfile()
clientProfile.httpProfile = httpProfile
clientProfile.signMethod = "TC3-HMAC-SHA256"
client = asr_client.AsrClient(cred, "ap-shanghai", clientProfile)
# 读取文件以及base64
with reader.SaveOpen(to_mp3(wav_path), 'rb') as fwave:
data = str(fwave.read())
dataLen = len(data)
base64Wav = base64.b64encode(data)
# 发送请求
req = models.SentenceRecognitionRequest()
params = {
"ProjectId": 0,
"SubServiceType": 2,
"EngSerViceType": "16k_en",
"SourceType": 1,
"Url": "",
"VoiceFormat": "mp3",
"UsrAudioKey": "session-123",
"Data": base64Wav,
"DataLen": dataLen
}
req._deserialize(params)
resp = client.SentenceRecognition(req)
if show_detail:
print(resp.to_json_string())
# windows系统使用下面一行替换上面一行
# print(resp.to_json_string().decode('UTF-8').encode('GBK') )
words = []
for w in resp.Result.split():
if not re.match('[a-zA-W]', w[-1:]):
w = w[0:-1]
words.append(w)
return ' '.join(words).lower()
except TencentCloudSDKException as err:
print(err)
return '[ERROR]'
def sentence_asr(self):
# 本地文件方式请求
try:
cred = credential.Credential(self.SecretId, self.SecretKey)
httpProfile = HttpProfile()
httpProfile.endpoint = "asr.tencentcloudapi.com"
clientProfile = ClientProfile()
clientProfile.httpProfile = httpProfile
clientProfile.signMethod = "TC3-HMAC-SHA256"
client = asr_client.AsrClient(cred, "ap-shanghai", clientProfile)
# 读取文件以及 base64
# 此处可以下载测试音频 https://asr-audio-1300466766.cos.ap-nanjing.myqcloud.com/test16k.wav
with open(self.data, "rb") as f:
content = base64.b64encode(f.read()).decode('utf-8')
# 发送请求
req = models.SentenceRecognitionRequest()
if self.type:
params = {
"ProjectId": 0,
"SubServiceType": 2,
"SourceType": 1,
"UsrAudioKey": "session-123"
}
else:
params = {
"ProjectId": 0,
"SubServiceType": 2,
"SourceType": 0,
"UsrAudioKey": "session-123"
}
req.url = self.data
req._deserialize(params)
req.EngSerViceType = "16k_zh"
req.VoiceFormat = "wav"
if self.type:
req.DataLen = len(content)
req.Data = content
resp = client.SentenceRecognition(req)
res = resp.to_json_string()
res = json.loads(res)['Result']
res = res.replace('。', '').replace('，', '').replace(',', '')
return res
except TencentCloudSDKException as err:
print('tencent返回错误:%s' % err)
def recognize(self, path: PathLike, lang: str = ""):
if not lang:
lang = self.lang
if not isinstance(path, str):
return ["ERROR!", "File must be a path string."]
if lang not in self.lang_list:
return ["ERROR!", "Invalid language."]
try:
with BytesIO() as f:
f.seek(0)
audio = pydub.AudioSegment.from_file(path)\
.set_frame_rate(16000)\
.set_channels(1)
audio.export(f, format="wav", codec="s16le", bitrate='16k')
# f.seek(0)
# import pdb; pdb.set_trace()
data = f.getvalue()
data_len = len(data)
# print(data_len)
base64_wav = base64.b64encode(data).decode()
req = models.SentenceRecognitionRequest()
params = {
"ProjectId": 0,
"SubServiceType": 2,
"EngSerViceType": self.languages[lang],
"SourceType": 1,
"Url": "",
"VoiceFormat": "wav",
"UsrAudioKey": "catbaron.voice_recog",
"Data": base64_wav,
"DataLen": data_len
}
req._deserialize(params)
resp = self.client.SentenceRecognition(req)
# print(resp.to_json_string())
return [resp.Result]
except TencentCloudSDKException as err:
return ["ERROR!", str(err)]
#https://cloud.tencent.com/product/asr/getting-started
cred = credential.Credential("AKIDX1GGzR5U9avzoPSmfKl79uNaaQ17nK2T", "IJ9l1D9oBH7oHWhbEaeS6VK4XjdKkz0t")
httpProfile = HttpProfile()
httpProfile.endpoint = "asr.tencentcloudapi.com"
clientProfile = ClientProfile()
clientProfile.httpProfile = httpProfile
clientProfile.signMethod = "TC3-HMAC-SHA256"
client = asr_client.AsrClient(cred, "ap-shanghai", clientProfile)
#读取文件以及 base64
#此处可以下载测试音频 https://asr-audio-1300466766.cos.ap-nanjing.myqcloud.com/test16k.wav
with open('./12.mp3', "rb") as f:
if sys.version_info[0] == 2:
content = base64.b64encode(f.read())
else:
content = base64.b64encode(f.read()).decode('utf-8')
#发送请求
req = models.SentenceRecognitionRequest()
params = {"ProjectId":0,"SubServiceType":2,"SourceType":1,"UsrAudioKey":"session-123", "RequestId": '1c8400f1-913f-4938-a45f-0421b08782ad' }
req._deserialize(params)
req.DataLen = len(content)
req.Data = content
req.EngSerViceType = "16k_zh"
req.VoiceFormat = "wav"
resp = client.SentenceRecognition(req)
print(resp.to_json_string())
#windows 如果是 GBK 编码则用下面 print 语句替换上面 print 语句
#print(resp.to_json_string().decode('UTF-8').encode('GBK') )
except TencentCloudSDKException as err:
print(err,)
